To the Editor:

On the basis of human leukocyte antigen (HLA) association studies, workers in the field of psoriasis have long been aware that the HLA complex plays an important role in determining psoriasis susceptibility. The question has always been why many families appear not to show linkage to HLA. We share in the pleasure of Drs. Leder and Hodge (Leder and Hodge, 1999 [in this issue]) now that the genetics of the HLA region in psoriasis is coming into sharper focus.

The general agreement between Leder and Hodge's studies (Leder et al. Leder et al., 1998), our own work (Jenisch et al. Jenisch et al., 1998 and Jenisch et al., in press), and the recent studies of Trembath et al. (Trembath et al., 1997) and Burden et al. (Burden et al., 1998) provides welcome insight into this long-standing puzzle. By optimizing LOD scores over a variety of penetrance functions, assuming Hardy-Weinberg equilibrium, Leder and Hodge (Leder et al. Leder et al., 1998) found the highest LOD scores for dominant models specifying high disease allele frequency and low penetrance. We reached essentially the same conclusion, following the suggestions of Risch et al. (Risch et al., 1989) for complex-trait data. It is well appreciated that power to detect linkage is diminished when the disease allele frequency is high and the penetrance is low. Given the smaller sample sizes of earlier studies, it is not surprising that linkage to the HLA region was not always apparent.

We have reported that linkage to HLA is more readily detected when marker-trait disequilibrium is taken into account, in part because of more-accurate specification of phase (Jenisch et al. Jenisch et al., 1998). Leder et al. (Leder et al., 1998) and Trembath et al. (Trembath et al., 1997) report similar results. This effect was first pointed out 15 years ago (Clerget-Darpoux Clerget-Darpoux, 1982) but has not been widely exploited in the genetic analysis of other common HLA-associated disorders. Even without incorporation of disease-marker haplotype frequencies, Leder et al. (Leder et al., 1998) found strong evidence for linkage to HLA under a dominant model, whereas we did not. Leder et al.'s study made use of previously published pedigrees, and concerns regarding ascertainment bias in favor of linkage are inevitable in such a study. However, it is also possible that our sample yielded lower LOD scores because it contained a number of small pedigrees, thereby increasing the number of phase-unknown individuals.

We concur with Leder and Hodge (Leder and Hodge, 1999) that there is now excellent agreement regarding the importance of the HLA region in familial psoriasis and that this locus should now be referred to as PSORS1. We would emphasize that, because the HLA loci yielding the highest LOD scores in familial psoriasis are so similar to those observed in prior case-control association studies, there is unlikely to be any difference between familial and “sporadic” juvenile-onset psoriasis with respect to the involvement of PSORS1. We can also infer that genetic differences between juvenile- and adult-onset psoriasis must exist, because of their different HLA associations (Henseler and Christophers Henseler and Christophers, 1985). Whether an HLA locus different from PSORS1 is involved in the adult-onset form of this disease remains to be determined.

High disease allele frequencies and low penetrance values are likely to be the rule rather than the exception in common multifactorial diseases. We hope that these recent insights into the genetics of the HLA region in psoriasis will be of benefit to other groups studying complex genetic disorders.